Messenger RNA Fluctuations and Regulatory RNAs Shape the Dynamics of Negative Feedback Loop

Single cell experiments of simple regulatory networks can markedly differ from cell population experiments. Such differences arise from stochastic events in individual cells that are averaged out in cell populations. For instance, while individual cells may show sustained oscillations in the concentrations of some proteins, such oscillations may appear damped in the population average. In this paper we investigate the role of RNA stochastic fluctuations as a leading force to produce a sustained excitatory behavior at the single cell level. Opposed to some previous models, we build a fully stochastic model of a negative feedback loop that explicitly takes into account the RNA stochastic dynamics. We find that messenger RNA random fluctuations can be amplified during translation and produce sustained pulses of protein expression. Motivated by the recent appreciation of the importance of non–coding regulatory RNAs in post–transcription regulation, we also consider the possibility that a regulatory RNA transcript could bind to the messenger RNA and repress translation. Our findings show that the regulatory transcript helps reduce gene expression variability both at the single cell level and at the cell population level.

💡 Research Summary

The paper investigates how stochastic fluctuations at the messenger RNA (mRNA) level influence the dynamics of a simple negative feedback loop (NFL) and how regulatory non‑coding RNAs can modulate this behavior. The authors construct a fully stochastic model that explicitly includes transcription, mRNA degradation, translation, protein degradation, and feedback inhibition of transcription by the protein product. Using Gillespie’s direct method, they simulate the system as a set of discrete reaction events rather than relying on deterministic ordinary differential equations.

Their simulations reveal that when the average number of mRNA molecules per cell is low (on the order of 10–30), transcriptional noise is substantial. Because translation is often highly efficient—each mRNA can generate dozens of protein copies in a short time—these mRNA fluctuations become amplified into pronounced protein “pulses.” In a single cell, these pulses appear as sustained oscillations, but the phases of the pulses differ across cells, so the population‑averaged signal looks damped. This demonstrates that apparent population‑level damping can mask robust, stochastic oscillations at the single‑cell level.

To explore the role of post‑transcriptional regulation, the authors add a regulatory RNA (R) that can bind to the mRNA, forming an inactive complex that blocks translation and is eventually degraded. By varying the binding, unbinding, and degradation rates of the R‑mRNA complex, they show that sufficient expression of the regulatory RNA dramatically reduces the number of free mRNA molecules, thereby lowering the amplitude and frequency of protein pulses. Quantitatively, the coefficient of variation (CV) of protein levels drops by 30–40 % in single cells, and the population average converges to a stable fixed point with minimal oscillatory remnants.

Parameter sweeps indicate that the impact of the regulatory RNA is most pronounced when the feedback strength is weak; under strong feedback the system already suppresses fluctuations, making the RNA’s contribution marginal. Conversely, when translation efficiency is high, the RNA’s ability to sequester mRNA becomes a critical noise‑damping mechanism. The authors also perform sensitivity analyses that confirm the robustness of these findings across a broad range of transcription and translation rates.

Overall, the study provides three key insights: (1) mRNA stochasticity can be a primary driver of sustained excitatory dynamics in simple NFLs; (2) translation can act as an amplifier of upstream noise, converting modest mRNA fluctuations into large protein bursts; and (3) non‑coding regulatory RNAs serve as effective post‑transcriptional buffers that reduce both single‑cell variability and population‑level oscillations. The work bridges a gap between theoretical models that often ignore RNA‑level noise and experimental observations of heterogeneous single‑cell behavior, suggesting that incorporating RNA dynamics and regulatory RNAs is essential for accurate modeling of gene‑regulatory circuits. Future experimental validation and extension to more complex networks are proposed as natural next steps.